tracing-root-causeslisted

# Tracing root causes <!-- Methodology distilled from OMC's tracer agent (MIT), 2026-07-30. See docs/adrs/ADR-002-omc-salvage.md for provenance. --> Explain outcomes through evidence, not narrative. The failure mode this prevents: jumping from symptom to favorite explanation, then collecting only confirming evidence. ## The discipline 1. **Observe before interpreting.** Restate exactly what was observed — which artifact, what behavior, when. If you catch yourself rewriting the observation to fit a theory, stop. 2. **Compete the hypotheses.** Under ambiguity, hold ≥2 explanations from deliberately different frames: code path, config/environment, measurement artifact, external dependency, timing. "The measurement is wrong" is always a candidate. 3. **Rank evidence by strength.** From strongest to weakest: controlled repro / discriminating experiment → primary artifact with tight provenance (timestamped logs, git history, file:line behavior) → independent sources converging → single-source inference that fits → circumstantial (naming, temporal proximity, stack position) → intuition. When tiers conflict, the higher tier wins; never treat support as flat. 4. **Seek disconfirmation.** For each serious hypothesis ask: "what should we observe if this were true — do we?" and "what observation would be hard to explain if this were true?". A hypothesis that survives only because nobody looked for counter-evidence keeps LOW confidence. 5. *